Research Agents

Documentation Index

Fetch the complete documentation index at: https://docs.getclaro.ai/llms.txt

Use this file to discover all available pages before exploring further.

Research Agents are fast, task-specific AI agents for ad-hoc work that doesn’t (yet) belong in your persistent catalogue. They generate lists, extract structured data from documents, enrich spreadsheets, and scrape URLs — with no schema setup required. If the result of an agent run is something you want to maintain over time, promote it into a Catalogue.

---

​

The four core agents

​

Find your perfect list

Describe your ideal prospects, suppliers, or partners and Claro generates a researched dataset with key attributes.

• Input — a natural-language brief plus any seed criteria (region, size, industry, etc.).
• Output — a typed dataset with rows, attributes, and per-cell citations.
• Use — outbound prospecting, supplier discovery, market mapping.

​

Turn documents into structured data

Extract structured records from one or many documents at once.

• Input — PDFs, scanned docs, contracts, datasheets, or reports plus a target schema (or infer).
• Output — a typed dataset, one row per source document (or per logical entity), with per-field provenance.
• Use — invoice batches, datasheet libraries, contract analysis, compliance ingest.

​

Analyze & enrich spreadsheets

Upload an existing spreadsheet and Claro validates, standardizes, and completes missing data.

• Input — CSV / XLSX, plus an enrichment goal (fill missing, validate against rules, add derived columns).
• Output — the enriched sheet with confidence scores per cell.
• Use — quick clean-up of supplier exports, lead lists, or partner data before further work.

​

Scrape data from URLs

Collect structured records from web pages.

• Input — a list of URLs (or a base URL with crawl rules) plus a target schema.
• Output — a typed dataset with per-record source URL.
• Use — competitor catalog snapshots, content audits, location data collection.

---

​

Where outputs live

• Generated Datasets — every agent run produces a dataset listed here.
• Uploaded Files — input files retained for re-use and reproducibility.

Both surfaces appear in the sidebar under Research Agents.

---

​

Promoting a dataset into a Catalogue

When a one-off result becomes ongoing, promote it. From any Generated Dataset:

1. Choose Promote to Catalogue.
2. Pick the target catalogue (existing or new).
3. Map columns to attributes. New attributes can be created on the fly.

After promotion the records become persistent and subject to all the same operations, provenance tracking, and review gates as native catalogues.

---

​

Agent Library

Research Agents is not a closed set. Browse Agent Library in the sidebar for additional task-specific agents (added regularly), and pin the ones your team uses most to your dashboard.

---

​

When to use a Research Agent vs. a Catalogue operation

The two surfaces complement each other. Many teams use Research Agents to scope a dataset, then promote it into a catalogue once the schema and quality bar are clear.